Alexa Skills — today and future
Advancements in speech technology over the last 2 years have led to rapid advancements in hardware and software products launched in the space.
Where Siri (Apple) enjoyed monopoly earlier, Alexa (Echo) and Google Voice Assistant (Google Home) have been strong contenders, despite being relatively late entrants. Alexa has significantly moved the goal post for Siri. The past few months have also witnessed announcements from Microsoft, IBM, and Google, all hitting new milestones in speech recognition accuracy. The claimed error rate has reduced to 5.5% now — which is marginally lower than what humans do (reference)!
Alexa is leading the way in the voice assistant space with expected sales in 2017 to be 25M units. Alexa has also introduced a platform capability to the last generation of which would enable a faster and richer growth.
Leveraging this emerging platform, I have been thinking about what kind of Alexa apps would be uniquely positioned, to be better created out of India? For the Indian market consumption? For global market consumption? Can we build products that ride on the Alexa adoption wave and get a head-start? New platforms give a chance for new networks to be created and new solutions to emerge. This can be a golden opportunity to create a large company — the company of the decade — riding on the Alexa platform.
Use cases of Alexa:
As a consumer, Alexa is clearly better than anything out there for at least 2 known use cases: 1) note taking and 2) playing music
But there are many other use cases that have also been developed and are being used. I have collated this list of use-cases from a few sources. My intent is to brainstorm on what can be created out of India.
Consumer use cases:
1) Entertainment: Consume content like jokes, Trivia related to Tonight Show/Game of Thrones, bed-time stories, songs, games like Jeoprady, runescape, famous quotes, podcasts, audiobooks, etc
2) News, Traffic, Weather and Informational Skills: Provide news headlines, traffic information, weather, stock quotes, food recipes, search in wikipedia, etc
3) Health and Wellness Skills: Assist in meditation, Assist in exercises, Give health info, medicine reminders, Book appointment, Keep track of vitals, etc.
4) Clock skills: Set alarm/timer
5) Smart Home Skills: Control garage doors, door bells, lights, TVs, fitbit, thermostat, etc
6) Note-taking Skills: Add / edit items in to-do-list, Add reminders, etc
7) Purchase related Skills: Order items, Track packages, Order Uber/Lyft, Check for deals, Book flights, etc
Enterprise use cases:
1) Meeting Notes and Transcription related Skills — Jotting down notes during the meeting, recording Doctor conversation and prescription transcription are the most obvious use cases.
2) Sales Feedback related Skills — A growing number of sales interactions are taking place over the phone, and sales executives are concerned about becoming blind to these conversations. Chorus.ai and Gong.ai leading the way for startups enabling better sales training, sales guidance for more efficiency.
This is just a dump of use cases. The intent is to get the conversation started and connect with like-minded folks.
Features of Alexa:
The three power features in mobile world which enabled mobile first/native companies were a) always-with-you, b) location and c)camera. I believe that Alexa is a powerful due to the following features:
1) Low tech: Easy to use for kids and older people. The first time and not so tech savvy users can use other tech heavy services using Alexa much easier
2) Stack of actions and state sensitivity: Alexa can handle multiple commands by stacking them, run multiple programs in parallel and understand which command would be applied to which program with ease. This might seem small but imagine giving a phone the command to wake me up after 30 minutes, play soothing music till then, interrupt music and ring the alarm if I get an email from my boss, and after 15 minutes, give new command to pause music and cancel the alarm.
3) Always listening and hands-free: This is more important than it seems. Unlike mobile phones where all actions other than location is to be specifically given, in Alexa, the default mode is always listening. Its hands free — so its applicable in many more scenarios than mobile would be. Alexa knows more about you than your mobile phone would.
4) Shared in group: This is the first shared device at large scale. Family members and employees can have shared to-do lists, meeting minutes, communicate asynchronously, etc
I admittedly do not know enough. Best entrepreneurs will certainly have more refined thoughts.
Some unstructured suggestions from what I have learnt from various entrepreneurs:
- Since the technology is not very evolved, building a general purpose AI solution does not make sense. The first few products would probably be very vertical focused so that the problem is more constrained and solvable.
- Because I am thinking more vertical solution with focused approach, intuitively the first problems that I see getting solved are enterprise use cases like healthcare doctor practice management, sales enablement, meeting transcription, etc. Easier to distribute and sell in those markets. But I might be wrong here. For enterprise use cases, constraints of understanding the universe don’t apply. They are by definition constraint to certain database for that vertical/use case and it starts with a default context. The value will be, if one could parse the transcript, extract facts, index them and provide query interface. A very hard engineering problem, but doable to some extend because of the constraint. It can be done in sales organisation, hospitals, schools, restaurants, government services, marketing, HR, etc. Going vertical focused keeps the problem in check.
- Think of Alexa as just a speech to text engine and a distribution platform. Think about “long term” sustainable local advantages. Some ideas I come across seem good for now but would be eventually taken away by global players if your technology is not globally competitive.
If you’re an entrepreneur using voice, I would be happy to be your sparring partner and brainstorm. Lets build awesome companies together :-)
Disclaimer: Views expressed here are my personal reflections and not indicative of the views of my current or previous employers. No part of this post maybe reproduced or quoted without explicit permission.
